Embed - Moozonian Search

news.ycombinator.com• May 26, 2026• 1 min read

Why codex /goal fails on complex workflows: compaction amnesia and context rotHi HN,When Openai released `/goal` earlier this month, I was really excited to try it for long-horizon tasks. But after using it, it didn't blow me away and i did some digging and found a major architectural flaw when using it for complex multi-issue workflows: context rot.This isn't anything new, but given how openai positioned this feature to developers, i was let down by how they'd implemented context management.Though /goal is a step forward in long-horizon coding, it lacks task decomposition and proper handling of context - it uses a multi-tier approach that includes persistent context chaining (PCC) to memory, local vector embeddings for RAG, sliding windows, and compaction.In principle, giving codex a directive of `/goal work towards closing my open issues on github` should work but this specific execution model hits a fatal wall - Even with massive context windows and RAG, llm reasoning quality degrades significantly beyond 100-150k~ tokens, the agent continues working with wor

bing.com• May 25, 2026• 1 min read

Zepp Health (ZEPP) Q4 2025 Earnings TranscriptBrand strategy incorporated “structure level exposure” via HYROX event integration, with sponsorship visibility embedded directly into athletic results rather than traditional advertising. The company ...

github.com• May 21, 2026• 1 min read

Show HN: Darc – grep-like memory search tool for coding agentsHi HN, I’m Junha Park. I've been experimenting with agent memory, especially how to make agents run more reliably on large tasks.I built Darc, an open-source shared memory search tool for coding agents, with a different approach from most agent memory systems we can see today. It is an index + (lexical) search tool over agent session history, rather than a managed memory system. No embeddings, no agent-aided consolidation, no injection hooks.How it works: Darc archives Codex / Claude Code session rollouts that already exist under ~/.codex and ~/.claude, indexes them into a single SQLite DB, and exposes session/turn/tool call/file-level search commands over them.Why I built this: I've tried using different agent memory tools and they were useful, but sometimes I found them nosiy. I kept turning memory on/off depending on the task. For example, when I ran iterative code review rounds, I saw reviewer agents report "no findings" with "memory cited" memos, where the cited memory came from r

apps.apple.com• May 19, 2026• 1 min read

Sieve – scans Cursor/Claude chat history for leaked API keysBackground: I was using Cursor to set up an OpenAI integration.The agent read my .env file, added the key to the config, and everything worked. What I didn't think about: that key was now sitting in a plaintext SQLite database at ~/Library/ApplicationSupport/Cursor/User/workspaceStorage/..AI coding tools (Cursor, Claude Code, Copilot, Cline) routinely read .env files as part of normal operation. Every secret they touch gets embedded in their local transcript/state files — unencrypted, outside .gitignore, persisted indefinitely.Standard secret scanners (gitleaks, detect-secrets) scan git repos. Nobody scans AI transcript stores. That's the gap.Sieve scans those files locally on your Mac. Flags exposed keys by severity. Redacts them in-place. Stores fingerprints in Keychain — never plaintext. Covers Cursor, Claude Code, Claude Desktop, Copilot, Cline, Roo Cline, Windsurf, Gemini CLI, and .env files.Happy to answer questions about how the SQLite parsing works or the detection rules.

github.com• May 17, 2026• 1 min read

Show HN: Semble – Code search for agents that uses 98% fewer tokens than grepHey HN! We (Stephan and Thomas) recently open-sourced Semble. We kept running into the same problem while using Claude Code on large codebases: when the agent can't find something directly, it falls back to grep, reading full files or launching subagents. This uses a lot of tokens, and often still misses the relevant code. There are existing tools for this, but they were either too slow to index on demand, needed API keys, or had poor retrieval quality.Semble is our solution for this. It combines static Model2Vec embeddings (using our latest static model: potion-code-16M) with BM25, fused via RRF and reranked with code-aware signals. Everything runs on CPU since there's no transformers involved. On our benchmark of ~1250 query/document pairs across 63 repos and 19 languages, it uses 98% fewer tokens than grep+read and reaches 99% of the retrieval quality of a 137M-parameter code-trained transformer, while being ~200x faster.Main features:- Token-efficient: 98% fewer tokens than grep+re

infiniteswap.app• May 16, 2026• 1 min read

Show HN: Infinite Swap – Trade a bottle cap up to a houseHey HN,Infinite Swap is a game inspired by Kyle MacDonald’s One Red Paperclip and other bigger-and-better swap stories - you start with a green bottle cap and trade it with other characters in various locations all the way up to a house and beyond.A while ago I got inspired by viral games Infinite Craft and What Beats Rock? and wanted to make my own ‘exploring a LLM’s latent space’ style challenge. After much experimentation this became a trading style game, where the LLM comes up with all the items, trading scenarios and value estimates, which ended up being surprisingly fun and somewhat of a roller coaster to play.I found it was the most intriguing if I separated the item offer from its valuation - so I implemented a pricing oracle called “Sal” who gives this info with his, sometimes snarky, opinion after each trade.The item offer mechanism goes deeper too - each character’s profile is combined with a set of hidden motivations. I then use embeddings and cosine similarity to represent

hypercubic.ai• May 12, 2026• 1 min read

Show HN: Agentic interface for mainframes and COBOLHi HN, we’re Sai and Aayush, and we’re building Hypercubic (https://www.hypercubic.ai/), bringing AI tools to the mainframe and COBOL world. (We did a Launch HN last year: https://news.ycombinator.com/item?id=45877517.) Today we’re launching Hopper, an agentic development environment for mainframes.You can download it here: https://www.hypercubic.ai/hopper, and you can also request access and immediately get a mainframe user account to play with.There's also a video runthrough at https://www.youtube.com/watch?v=q81L5DcfBvE.Mainframes still run a surprising amount of critical infrastructure: banking, payments, insurance, airlines, government programs, logistics, and core operations at large institutions. Many of these systems are decades old, but they continue to process enormous transaction volumes because they are reliable, secure, and deeply embedded into business operations.A lot of that software is written in COBOL and runs on IBM z/OS. The development environment looks very differ

news.ycombinator.com• May 8, 2026• 1 min read

Show HN: Postlet – Tiny static blog generator with plugins and themes supportKeeping it aligned with my philosophy of simplicity, extensibility, and usefulness, I built this.Github: https://github.com/freakynit/postletProject homepage (built using postlet itself): https://postlet.pagey.site/My personal blog (also built using postlet): https://nb1t.sh/--Full plugin based architecture. Root or subpath deployment. Page tagging, SEO meta, heading anchors. Theme support plus per-page CSS overrides. Automated bundling of other assets. Automated related posts enrichment using vector embeddings or local keyword scoring.Entire code-base (excluding css), including comments is less than 1000 lines.

github.com• May 7, 2026• 1 min read

Show HN: wfb-link, a userspace WiFiBroadcast radio stack for macOSHi HN, I’ve been working on a Rust userspace radio stack for running WFB-style links from macOS using RTL8812AU USB adapters. Full disclosure: I'm a software engineer, but not really a hardware or embedded systems engineer, so Codex GPT 5.5 has done the lion-share of the work here along with a bit of help from Claude Opus 4.7 here and there. It's taken about 1.5 to 2 weeks to get from zero to this first release.macOS doesn't expose the monitor-mode / packet-injection path that WFB systems normally rely on. I really didn't like the idea of needing a separate linux box just to talk WFB to other edge devices, like drones. This talks to the ALFA AWUS036ACH as a USB peripheral directly, initializes the RTL8812AU, submits raw 802.11 WFB frames over bulk OUT, receives frames over bulk IN, and bridges them to WFB-NG’s distributor/aggregator UDP protocols.Basically, this is what's working (you can see more detail in the readme):- native macOS userspace RTL8812AU bring-up- TX/RX of WFB datagrams

news.ycombinator.com• May 7, 2026• 1 min read

Ask HN: Does human approval exist in your agent stack? why/why not?Long-time SaaS GTM guy with product fwd lens. New to infrastructure, shamelessly trying to learn. Go easy on me.Building on a thesis that human approval will ultimately need to be more embedded into meaningful human/agent workflow than fully autonomous (learning the hard way since our lobster friend entered the chat). The question I keep asking myself is "did I actually authorize ClaudeRod (my lobster) to do this". Recent news has me more concerned.I've been hacking on a solution but again, I'M NOT A DEV. I know how to recognize pain and chart a mental map to solution - I've done this for 20 yrs. But I don't know if I have enough genuine feedback yet to quantify the pain. Three patterns I see from research that I'd appreciate any/all genuine feedback on: 1. Confirm before it runs: Agent proposes, human authorizes (quick manual click), Agent executes. Seems logical from an audit trail, but kills flow. 2. Notify after: Agent acts with short window to 'undo', like gmail. Lower friction, b

dreamworkhq.com• May 7, 2026• 1 min read

Show HN: Dreamwork – a job search site I made after Indeed fired my pregnant wifHey, I’m Colin and I have a fun story for you.My 7 month pregnant wife was laid off from Indeed (she was a PM there) back in December. This pissed me off quite a bit, as she was supposed to get 6 months leave and instead got fired.So I spent the last five months working part time to build Dreamwork, a platform aiming to make the job search experience actually better with AI (not just mass application spam).I started with just a telegram bot doing scraping, then advanced to Google Embeddings 2.0 for vectorizing the jobs, built out a tight 6 axis scorecard for both the user and each job.Then I actually got to use my English degree (lol) to optimize the prompt for custom per-job resumes and cover letters to make them not sound like - again - shitty AI.Most AI cover letters have a kind of consistently dead quality. They use all the keywords and somehow communicate nothing. I absolutely hate that, so I’ve been fairly obsessive about making the output feel more like a decent human draft: spe

news.ycombinator.com• May 5, 2026• 1 min read

Ask HN: What tech hobbies are worth exploring?Software is being overwhelmed with AI-generated noise with no sign of slowing. This makes it increasingly difficult to find value and to feel valued.I want to feel the enthusiasm I felt when I first started programming, and I've starting looking into a hardware oriented tech hobby. Something involving a smallish group of dedicated people, minimal hype, working with cool tech, and maybe feeling like we're building the future.Some areas I'm considering include robotics, sensors/IoT and embedded systems.These are my constraints:- Can be done at home- Reasonable budget- Surmountable learning curve to a newcomer- Strong community, minimal hypeCan you recommend an area of tech that fits the above? Doesn't have to involve hardware, I just figure there's a higher barrier to entry and therefore less noise.

pulsepages.co• May 5, 2026• 1 min read

Show HN: PulsePages – Multi-page websites for $9/year (Carrd alternativeCarrd is good at one thing: one page. The moment you need a second page — a /pricing, an /about, a /blog — you're either hacking single-column scroll or jumping to Squarespace ($192/yr) or Wix ($192/yr+). Neither is proportionate for a portfolio, indie project, or small business that just needs a few pages.So I built PulsePages. Drag-drop block editor, unlimited pages, custom domains with auto-SSL, all for $9/year. Free tier covers unlimited single-page sites with no credit card.What's in it: - Block editor: text, images, video, embeds, contact forms - Custom domains + automatic SSL provisioning - AI writing assistant (drafts and rewrites page copy in context) - Version history: roll back any page to any saved state - Gated content: password-protect individual pages or sections - Free tier: unlimited single-page sites, no account required to tryStack: Next.js, Supabase, Stripe, Render, Resend.Live at pulsepages.co. Break things, leave feedback here or at pulsepages@polsia.app.

news.ycombinator.com• May 5, 2026• 1 min read

I'm looking for an AI Automation Engineer role or gigHi all, I'm an AI automation engineer who builds systems that replace manual work, scale outreach, and turn workflows into revenue. I have sent out working systems for managing leads to CRM, finding real estate deals, sorting emails with AI, chatbot integration, AI voice integration, and agentic RAG systems. I am also always improving my skills in AI agents and LLM integration & RAG embeddings.I’m looking for roles where I can build and deploy systems that eliminate manual work and scale revenue. I enjoy working with tools like n8n, Airtable, APIs, LLMs, and data-driven workflows to solve real operational bottlenecks. I prefer environments where automation is core to growth, not an afterthought. I value clear communication, fast iteration, and ownership. I avoid roles that are limited to repetitive execution without room to design, optimize, and innovate.Feel Free to send me an email at Judedivine9@gmail.com

allbsides.com• May 4, 2026• 1 min read

Show HN: I indexed 8,643 BSides talks across 227 chapters and 6 continentsHi HN,I'm Roland, and for the past few weeks, I've been building AllBSides — a directory of every BSides conference talk uploaded to YouTube. As of today, 8,643 talks from 5,927 speakers across 227 chapters in 68 countries. Combined runtime is 280 days. The transcripts come to about 60 million words.The archive came together in stages:1. Manually map every BSides chapter's YouTube channel 2. Pull every video and transcript from Supabase 3. Run each transcript through Haiku for tag extraction (tools, topics, difficulty, team, talk style, research method, and much more) 4. Run results through Sonnet for categorization and dedup 5. Final pass goes through Opus for verification 6. Do a manual verification - at one time, the pipeline showed over 16k AI suggestions for manual verification. Today, most are resolved.Total LLM cost so far: about €200. The whole pipeline is rebuildable from scratch.Each talk gets its own page with embedded video, full transcript, speakers, tags, and "related tal

toposonico.com• May 2, 2026• 1 min read

Show HN: A navigable map and recommender for 17M music entitiesHello HN,This is toposonico, a music recommender and navigable map. At core it's a skipgram word2vec model trained over ~6M playlists. Tracks are embedded in a 128d space. Embeddings for albums, artists and labels are computed marginalizing over tracks. The 2D map was built with UMAP.Both the model and UMAP were trained in the cloud over a NVIDIA A100. All things considered it cost me around ~50EUR, over two main training sessions and a few experiments. For the slippy map I experimented with a few libraries. Ended up with Maplibre GL JS. Loved working with it, kudos to their developers. For the recommender indexes I used FAISS, another fantastic piece of software. Pretty happy with the thing running on a small and cheap box.Two things influenced me in making this. The first is decade-old idea: human navigation and exploration skills work in information spaces too. Many ML concepts fit this idea especially well. It would be nice to see more experiments in this direction. The second goes

github.com• May 1, 2026• 1 min read

Show HN: An Agent First Slack CLIHey folks,My team and I have been building a background agents as a service product. One of the things we needed pretty early on was for some way for the agents to be able to drive slack.Right now, I don't think there are many good agent-first ways of doing this. I don't love MCP -- it's just too many tokens in the context window, and agents seem to do better with CLIs because they can embed them in code and so on. But other CLIs that exist are either a) subsets of the web api that only focus on human needs and b) have polished terminal UIs with spinners and colors and interactive things that make it hard for LLMs to actually use them.nori-slack-cli is a very thin dynamic wrapper over @slack/web-api (i.e. bolt). Whatever the SDK can call, the CLI can call. When slack adds a new method it should just work without a new release on our end. (Inspired by the GWS cli: https://github.com/googleworkspace/cli).So you(r agent) can do things like:`nori-slack chat.postMessage --channel C123 --tex

github.com• May 1, 2026• 1 min read

Show HN: Superkube - Rewriting Kubernetes in RustI have embarked on a journey of rewriting Kubernetes into a single binary in Rust, with everything embedded.Architecturally, instead of etcd, it has options to use SQLite or PostgreSQL as a backend. I am trying to make it compatible with Kubernetes APIs, and kubectl should just work out of the box.This is not 100% ready yet, but I just wanted to show the first look. It is almost 90% AI-generated. Claude code with Opus 4.7 is being used here.It should just work out of the box on macOS with Docker as runtime.